This analysis attempts to predict NBA performance based on college statistics. The sample contains college players drafted from 2011-2016. The data includes box score statistics, advanced college statistics, and physical measurements. The "Complete" model contains the data stated above as well as the location in which the players were drafted or predicted to be drafted. Unsurprisingly, I found this to improve the predictive ability of the model. NBA teams have reasonable predictive ability and can generally identify the best prospects. The "Stats Only" model uses only box score statistics, advanced college statistics, and physical measurements to predict NBA performance. It is useful to think about uncertainty in the framework of a probability distribution. I used machine learning techniques to predict the probability that an NBA prospect will be an all-star, starter, bench player, and bust. Included inhere are the out of sample predictions for the 2017 and 2018 Draft Class for both the Complete and Stats Only models.